Anthropic: Claude Sonnet 5
anthropic/claude-sonnet-5
Context: 1,000,000In: $2.00/1MOut: $10.00/1MUpdated 2026-10-04
Sonnet 5 is Anthropic's most capable Sonnet-class model, with frontier performance across coding, agents, and professional work. It supports adaptive thinking with selectable reasoning effort levels (low, medium, high, max,...
Intelligence
38.2
Coding
71.5
Agentic
43.6
Metric profile
Trend over time
Classic evals (curated from model cards)
Anthropic Claude — manually maintained, may lag releases
MMLU-Pro
74%
GPQA Diamond
70.8%
HumanEval
92%
Design Arena (ELO by category)
| Category | ELO | Win rate | Rank |
|---|---|---|---|
| gamedev | 1296 | 53% | #16 |
| uicomponent | 1286 | 54.1% | #22 |
| codecategories | 1283 | 53.1% | #18 |
| website | 1283 | 52.9% | #19 |
| 3d | 1277 | 54.4% | #21 |
| dataviz | 1258 | 52.9% | #27 |
| fullstack | 1248 | 53.7% | #9 |
| androidnative | 1237 | 54.4% | #9 |
| webapps | 1237 | 53.3% | #10 |
| agenticgamedev | 1226 | 52.3% | #7 |
| godotgamedev | 1225 | 54.4% | #13 |
| asciiart | 1221 | 52.1% | #19 |
| htmlslides | 1217 | 53.7% | #5 |
| mobileapps | 1216 | 51.5% | #11 |
| python-pptxslides | 1216 | 50.1% | #13 |
| svg | 1190 | 49.9% | #28 |